Modeling and Evaluation of the Performance of Parallel/Distributed File System

The mass data storage systems need to be coupled with efficient parallel/distributed file systems, such as Lustre and HDFS, which can effectively solve the problems of the mass data storage and I/O bottlenecks. This chapter systematically studies the performance factors and distribution of parallel/distributed file systems and proposes a valuation scheme for the classic parallel/distributed file system by capturing the changes in workload characteristics. The experiment results show that the proposed evaluation scheme can reach better accuracy and efficiency.

[1]  Jarek Nieplocha,et al.  Evaluation of active storage strategies for the lustre parallel file system , 2007, Proceedings of the 2007 ACM/IEEE Conference on Supercomputing (SC '07).

[2]  Depei Qian,et al.  A High Availability Mechanism for Parallel File System , 2005, APPT.

[3]  Qiang Cao,et al.  Approximate parameters analysis of a closed fork-join queue model in an object-based storage system , 2009, International Workshop on Information Data Storage and International Symposium on Optical Storage.

[4]  Jeffrey S. Vetter,et al.  Performance characterization and optimization of parallel I/O on the Cray XT , 2008, 2008 IEEE International Symposium on Parallel and Distributed Processing.

[5]  Wen-Zhan Song,et al.  Cooperative Resource Sharing and Pricing for Proactive Dynamic Spectrum Access via Nash Bargaining Solution , 2014, IEEE Transactions on Parallel and Distributed Systems.

[6]  Jeffrey S. Vetter,et al.  Empirical Analysis of a Large-Scale Hierarchical Storage System , 2008, Euro-Par.

[7]  Yinliang Yue,et al.  High Availability Storage System Based on Two-Level Metadata Management , 2007, 2007 Japan-China Joint Workshop on Frontier of Computer Science and Technology (FCST 2007).

[8]  Jacob Benesty,et al.  Pearson Correlation Coefficient , 2009 .

[9]  Verdi March,et al.  Evaluation of a Performance Model of Lustre File System , 2010, 2010 Fifth Annual ChinaGrid Conference.

[10]  Jeffrey S. Vetter,et al.  Exploiting Lustre File Joining for Effective Collective IO , 2007, Seventh IEEE International Symposium on Cluster Computing and the Grid (CCGrid '07).